一、hbase shell的基础命令
# 版本号查看
[root@Test-Hadoop-NN-01 hbase]$ ./bin/hbase version
HBase 2.4.0
Source code repository git://apurtell-ltm.internal.salesforce.com/Users/apurtell/src/hbase revision=282ab70012ae843af54a6779543ff20acbcbb629
# 客户端登录
[root@Cloud-Hadoop-NN-01 hbase]$ ./bin/hbase shell
# 查看所有表
hbase> list
### 创建数据表student，包含Sname、Ssex、Sage、Sdept、course列族/列
### 说明：列族不指定列名时，列族可以直接成为列名，从下面的数据插入可以看出 ！！！
hbase> create 'student','Sname','Ssex','Sage','Sdept','course'
# 描述表信息
hbase> describe 'student'
Table student is ENABLED
student
COLUMN FAMILIES DESCRIPTION
{NAME => 'Sage', BLOOMFILTER => 'ROW', IN_MEMORY => 'false', VERSIONS => '1', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', COMPRESSION => 'NONE', TTL => 'FOREVER', MIN_VERSIO
NS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}
{NAME => 'Sdept', BLOOMFILTER => 'ROW', IN_MEMORY => 'false', VERSIONS => '1', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', COMPRESSION => 'NONE', TTL => 'FOREVER', MIN_VERSI
ONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}
{NAME => 'Sname', BLOOMFILTER => 'ROW', IN_MEMORY => 'false', VERSIONS => '1', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', COMPRESSION => 'NONE', TTL => 'FOREVER', MIN_VERSI
ONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}
{NAME => 'Ssex', BLOOMFILTER => 'ROW', IN_MEMORY => 'false', VERSIONS => '1', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', COMPRESSION => 'NONE', TTL => 'FOREVER', MIN_VERSIO
NS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}
{NAME => 'course', BLOOMFILTER => 'ROW', IN_MEMORY => 'false', VERSIONS => '1', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', COMPRESSION => 'NONE', TTL => 'FOREVER', MIN_VERS
IONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}
5 row(s)
# 插入数据表student：Sname:xing='Shen'
hbase> put 'student','10001','Sname:xing','Shen'
hbase> put 'student','10001','Sname:ming','HY'
hbase> put 'student','10001','Ssex',1
# 查询数据表student：rowkey='10001'
hbase> get 'student','10001'
Took 0.0052 seconds
hbase:014:0> get 'student','10001'
COLUMN CELL
Sname:ming timestamp=2025-05-14T16:42:55.882, value=HY
Sname:xing timestamp=2025-05-14T16:42:43.429, value=Shen
Ssex: timestamp=2025-05-14T16:48:28.510, value=1
1 row(s)
Took 0.0161 seconds
# 查询数据表student：全部数据
hbase> scan 'student'
ROW COLUMN+CELL
10001 column=Sname:ming, timestamp=2025-05-14T16:42:55.882, value=HY
10001 column=Sname:xing, timestamp=2025-05-14T16:42:43.429, value=Shen
10001 column=Ssex:, timestamp=2025-05-14T16:48:28.510, value=1
1 row(s)
Took 0.0135 seconds
### 创建数据表scores：包含grade,course字段
hbase> create 'scores','grade','course'
# 插入数据表scores：rowkey=zhangsan01 ......
hbase> put 'scores','zhangsan01','grade:','101'
hbase> put 'scores','zhangsan01','course:art','90'
hbase> put 'scores','zhangsan01','course:math','99',1498003561726
#这里手动设置时间戳的时候一定不能大于当前系统时间，否则无法删除该数据。这里手动设置数据是为了下面的DependentColumnFilter过滤器试验。可以查看一下插入第一条数据的时间戳，再插入第二条数据的时间戳为第一条数据的时间戳
hbase> put 'scores','zhangsan02','grade:','102',1498003601365
hbase> put 'scores','zhangsan02','course:art','90'
hbase> put 'scores','zhangsan02','course:math','66',1498003561726
hbase> put 'scores','lisi01','grade:','201',1498003561726
hbase> put 'scores','lisi01','course:art','89'
hbase> put 'scores','lisi01','course:math','89',1498003561726
# 查询两个rowkey之间的数据
hbase> scan 'scores',{STARTROW=>'zhangsan01',STOPROW=>'zhangsan02'}
ROW COLUMN+CELL
zhangsan01 column=course:art, timestamp=1498003561726, value=90
zhangsan01 column=course:math, timestamp=1498003561726, value=99
zhangsan01 column=grade:, timestamp=1498003593575, value=101
1 row(s) in 0.0140 seconds
# 根据列名查询
hbase> scan 'scores',{COLUMNS=>'course:art'}
ROW COLUMN+CELL
lisi01 column=course:art, timestamp=1498003655021, value=89
zhangsan01 column=course:art, timestamp=1498003561726, value=90
zhangsan02 column=course:art, timestamp=1498003601365, value=90
3 row(s) in 0.0120 seconds
# 查询两个rowkey之间&根据列名来查询
hbase> scan 'scores',{COLUMNS=>'course:art',STARTROW=>'zhangsan01',STOPROW=>'zhangsan02'}
ROW COLUMN+CELL
zhangsan01 column=course:art, timestamp=1498003561726, value=90
1 row(s) in 0.0110 seconds
# 限制查找条数
hbase> scan 'scores',{LIMIT=>1}
ROW COLUMN+CELL
zhangsan01 column=course:art, timestamp=1498003561726, value=90
zhangsan01 column=course:math, timestamp=1498003561726, value=99
zhangsan01 column=grade:, timestamp=1498003593575, value=101
1 row(s) in 0.0140 seconds
# 查询时序数据
hbase> get 'pfme','dh_p_030_13030006@_@1747090800@_@'
# 查询时序数据
hbase> get 'pfme','dh_p_030_13030006@_@1747130400@_@FF-67-12-02-10-6A'
# 客户端退出
hbase> exit
二、hbase shell 实现rowkey的模糊查询
# 前缀匹配法
#查询以"2025"开头的rowkey，左闭右开区间（性能较高）
hbase> scan 'table_name', {STARTROW => '2025', STOPROW => '2025z'}
# 正则表达式过滤
#查询包含"123"的rowkey（性能较低，全表扫描时慎用，建议配合STARTROW/STOPROW缩小范围）
hbase> scan 'table_name', {FILTER => "RowFilter(=, 'regexstring:.*123.*')"}
# 子串匹配优化方案
#查询中间含"123"的rowkey（需预知固定长度），适用于已知字段固定位置时的精准匹配
hbase> scan 'table_name', { STARTROW => '000123', STOPROW => '999123z', FILTER => "ValueFilter(=, 'substring:123')"}
# 分页模糊查询
#分页查询含"123"的rowkey（每页10条），通过STARTROW参数实现翻页功能
hbase> scan 'table_name', { FILTER => "PageFilter(10) AND RowFilter(=, 'substring:123')"}
#eg: 查询DH专业的ct-tsdb
scan 'pfme', {STARTROW => 'dh_p_030_13030006@_@1747206000', STOPROW => 'dh_p_030_13030006@_@1747206000z'}
三、hbase shell 实现列值的模糊查询
#查询列族cf和列名c1中以123开头的所有行
hbase> scan 'table_name', {COLUMNS => ['cf:c1'], FILTER => "(PrefixFilter('123'))"}
#查询指定的列值中包含"123"的所有行
hbase> scan 'table_name', {COLUMNS => ['cf1:c1', 'cf2:c2'], FILTER => "ValueFilter(=, regexstring:'.*123.*')"}
参考文档：https://blog.csdn.net/m0_37739193/article/details/73615016#
